The HTML part of the message probably contains a tiny transparent image which is loaded from a URL that is different in each copy of the message. When your email client displays the message it requests the unique URL, telling the spammer that your email address is valid, as well as what IP address you're using (which can then be used to link your email address to info derived from web cookies etc).
Yahoo mail allows you to disable images in HTML emails for this reason, not sure about Hotmail, Outlook, Evolution etc.
I don't like to post to Slashdot without using at least one cliche and at least one acronym, so I should start by pointing out that IANAL.
But...
the downside of using a government agency for airport security rather than private companies is that the actions of government agencies fall under the equal protection clause of the 14th amendment. Far be it from me to suggest that a person's ethnic or religious background might be a good predictor of his or her tendency to hijack planes, but should that turn out to be the case, a government agency would not be able to stop people flying on the basis of a profile that included racial or religious information.
As far as I know there is nothing to stop a private company from refusing you service on any grounds it pleases, without giving a reason. If a gas station owner says "no gas for you, please leave my property" then as far as I know that's the end of the argument. No gas for you. So a private security company might be able to use information in its "profile of a terrorist" that the TSA would not be able to use for constitutional reasons. It would not be able to arrest people on the basis of the profile, but it would be able to refuse to let them through its checkpoints. (Of course there would have to be some kind of small print in the sales agreement to the effect that "the carrier or its partners (ie the security firm) may refuse to let you on the plane without giving a reason, and you will be given a refund instead" or suchlike.)
Re:Best examples of heresy I can think of
on
What You Can't Say
·
· Score: 1
I take your point that a rapist may be motivated by sex, but the statement "rape is about power" concerns the crime's meaning to the victim as well as the motivation of the rapist. To be raped is to experience a complete loss of control over your body, and the violation of the barrier between yourself and the outside world. (Not just physically but psychologically - I don't think it's too controversial to say that for most people, the difference between "Self" and "Other" is psychologically important. Maybe not to a Zen monk, but to everyone else.) That barrier is how we define ourselves, and if somebody else controls what is inside you and what is outside then you could feel that your sense of self has been destroyed. Destroying someone's sense of self is the strongest possible expression of power over them, short of killing them. So a particular rape might not be about power for the rapist, but I would argue that it is always about power for the victim.
Of course it's wrong to commit rape, along with any other violent crime. But whenever psycologists start talking, I'm skeptical. They have a theory and no hard evidence.
They may have been raped themselves, and they will probably have talked to a lot of victims of rape, so I suspect they know as much about the subject as you or I.
Actually, I'm skeptical anytime anyone starts talking about "root causes", not that I think it's inherently a bad way to think, but usually that ends up as some kind of "blame the parents of the parents of the parents of the parents of the murderer" argument.
I agree, it's sloppy thinking to confuse moral responsibility with causality. However, most penal systems aim to prevent future crimes as well as punishing past crimes, so understanding the causes of a crime is still important. If a serial rapist can be given psychological treatment that makes him less likely to commit another rape, shouldn't that be used in addition to punishment? It's not a matter of passing the buck to the rapist's parents or pedophile priests or anyone else, just a pragmatic matter of reducing the chances of future crimes.
I would guess it's mostly Sun or IBM workstations in universities and engineering companies, plus maybe a few obscure browsers/spiders that don't give enough information for Google to determine the OS.
Re:Best examples of heresy I can think of
on
What You Can't Say
·
· Score: 1
Rape is about sex, not power,
That's as stupid as saying that it's about power, not sex. By definition, both are involved.
and is a perfectly natural activity.
That has nothing to do with whether it's right. Saying that animals do something is not a moral justification for it.
Re:Best examples of heresy I can think of
on
What You Can't Say
·
· Score: 1
There's no reason to survive, but that doesn't mean it isn't worthwhile - reason isn't the only source of value.
Every reason depends on a set of axioms, and how do you choose the axioms? If you arrive at them by reasoning then you must have used other, deeper axioms. Eventually you have to face the fact that rationality has its limits - it can only build on your instinctive or ingrained emotional reactions, not define them. You can build wonderful structures with logic and reason, but the foundations will always be emotional.
Personally I found this realization quite liberating - I no longer felt the need to justify my feelings rationally - but it also killed any hope I had of understanding my place in the universe on an intellectual level. So what's left? You can still hope to understand the world on an intuitive level, and to reason correctly from that intuitive foundation. Or you can kill yourself - but there's no reason to do that either.;-)
Although I agree with your point, you referred to Europeans as "snotty" twice and "snobby" twice - I wonder if you're harbouring some stereotypes of your own?;-)
Don't like a company? Don't work for it, and don't buy from it. If enough people see things your way, it will go out of buisness. (And if enough people don't agree with you, it will stay in buisness - such is life in a dollar-vote based system.)
Gosh, I can't believe I never understood the elegant simplicity of capitalism before. All those millions of people in workers' movements around the world must be idiots - thank goodness you were here to introduce them to the simple truth that money is (and should be) power.
Don't like a company? Don't work for it
For most people there's no option - if work is available they have to take it, otherwise their children go hungry and they lose their home. If you live in a country where the welfare state provides you with the option of not taking a job because you don't like the company, it's because workers in previous generations fought against capitalists like you to create that welfare state.
such is life in a dollar-vote based system
Do you honestly think it's a good idea that political power should be based on wealth? Should one person be born with millions of times more "dollar-votes" than another? Or should we do whatever we can to limit the political influence of money and ensure that each human being, regardless of his or her income, is entitled to an equal vote? One of these principles is called capitalism, and the other is called democracy. Make your choice.
There is a big difference between creating a plan, and giving serious consideration to executing it.
True, and that's why this story is interesting - Britain believed the US was giving serious consideration to the use of force, not just planning for every contingency.
Most of the September 11 hijackers were from Saudi Arabia, but they certainly weren't funded by the Saudi government - Al Qaeda is committed to the overthrow of the Saudi royal family. If the US government avoids talking about Saudi Arabia it's not because their money supported the hijackers, but because their support for a fundamentalist dictatorship in the Middle East doesn't fit well with their supposed goal of democratizing the region.
An article at BigBlueBall.com states that 75% of web connections do not use a browser.
No it doesn't. The article says that "76% of active web surfers" use non-browser applications. It doesn't say they don't use a browser. Admittedly the headline implies that, but the phrase "web surfers" indicates the opposite.
What would "75% of web connections do not use a browser" even mean? Do you think 75% of people are surfing the web with wget?
you can't apply engineering rules to a human society. That's why we have sociologists, lawyers and politicians.
Perhaps we'll never be completely rid of them, but with the correct application of technology we could significantly reduce their numbers. Why do you think engineers get so excited about the possibility of sending people to Mars?
The government is now a constitutional democracy, so why is there so much homage paid to the archaic traditions and figureheads of the past?
First, the British government isn't constitutional in the same sense as the US government - there's no single document called "the British constitution". The founders of the US followed the European rationalist tradition: decide how the country should be run, write it down and embalm it for all time. (Until you change your mind - France has had five constitutions in 200 years.) In contrast, Britain's constitution follows the empirical tradition: if it ain't broke, don't fix it; when it breaks, patch it. So the British constitution is a messy tangle of legislation, common law and long-standing conventions, developed over time in a piecemeal fashion. Sort of a "release early, release often" approach to constitutional law. If the British constitution is Linux then the US constitution is Mach. (And the Magna Carta is Unix, the European Convention on Human Rights is the BSD networking stack, and the Copyright, Designs and Patents Act was written by SCO. Enough of that analogy.)
The book Systemantics, reviewed on Slashdot recently, claims that loosely-coupled systems developed in a piecemeal fashion are more stable than well-designed, tightly-coupled systems. I don't know if that's true of constitutions, but Britain has had a relatively peaceful (if slow) development from feudalism to near-democracy. Compared with almost any other country on Earth that's remarkably stable - even Belgium had a revolution.
Second, I think you're wide of the mark when you say that homage is paid to archaic traditions. British people are (in my experience) rather skeptical and cynical compared to Americans. If we tolerate archaic institutions it probably has more to do with suspicion of anyone who wants to rebuild the country in his own image (*cough*Blair*cough*) than with veneration of the past. When I visit the US I'm struck by the number of flags on display and the generally jingoistic atmosphere (and not just in the last two years). Many people seem to treat the US constitution as a sacred text, so I wonder whether there isn't more homage paid to archaic institutions in the US than in Britain (although the institutions are somewhat less archaic).
Most constitutions guaranteeing free speech and elections are as informative about the societies they allegedly define as a man saying 'Good morning' is about the weather.
I'm not really sure where I stand there. Like I said, due process plays a huge part in this. For example, if tickets are automatically doled out for speeding, do you have a strong method for appeal?
As a geek I'm worried by the increasing use of technological evidence in court cases. Geeks understand that technology fails, and more importantly that it can be hacked - information is not necessarily complete or correct just because it came from a computer or a camera. But most people attribute more significance to technological evidence than human evidence. Admittedly, cameras and computers don't lie of their own volition, but they can be made to lie by their operators. Evidence from a computer is only as valid as the word of the sysadmin, but is a jury more likely to believe a greasy hacker or a pristine computer printout?
Even if the information has not been tampered with, the use of high technology lends false authority to the entire chain of evidence. For example, it's well known that the odds of a false positive in a DNA comparison are billions to one. In the minds of many jurors, this will translate as "the odds that the accused is not guilty are billions to one". But in fact the odds are determined by multiplying all the probabilities in the chain of evidence - for example, what are the odds that a police officer substituted a DNA sample taken from the crime scene with one taken from the defendant? What are the odds that the defendant left DNA at the crime scene at some other time? What are the odds that the lab mixed up two samples? These are all much more likely than an error in the test itself, but because of our inability to estimate odds the strong link in the chain seems to lend strength to the weaker links, and the evidence seems unassailable.
To get back to surveillance, the best take I've heard on this issue came from a researcher in wearable computing (sorry, I've lost the link) who suggested that since we can't stop those in power from using surveillance technology, our best hope is to spread it as widely as possible (physically as well as socially). If everyone has a video camera and uses it all the time, it becomes hard to forge or misrepresent video evidence because you never know what conflicting evidence might exist. Likewise if everyone can find out where you've driven you have less to fear than if only the FBI can find out that information, because there's less chance that they'll try to misuse it.
I had no idea traffic shaping in Linux was so sophisticated. Looks like you can use setsockopt() to set the socket's priority via the TOS field. Thanks for the link!
You should be able to determine the transmitter's location with a single receiver, assuming that the transmitter can't teleport. But you have to do a lot of work in advance...
First, divide the building up into zones. Where possible, the boundaries between zones should follow "natural" boundaries that are likely to attenuate radio signals, like walls and ceilings, but if you have to cover large open spaces then you might have to draw some fairly arbitrary boundaries.
Put your receiver in the first zone and visit each of the other zones with your transmitter, recording the signal strength from each one. Then move the receiver to the second zone and repeat, and so on. Normalize the data - different hardware will produce different signals, so we're only interested in relative levels. For N zones, this gives you an NxN table of relative signal strengths between zones.
When you want to determine the location of a transmitter, first check whether the signal strength varies significantly over time. If it does, the source is probably moving. Stay in your current location (let's call it zone M) and make frequent measurements. By assuming that consecutive measurements come from the same zone or neighbouring zones (no teleporting), you can use the Mth row of your table to guess the source's location.
If the transmitter's signal strength remains constant it is probably still, so move around the building and take measurements from several zones. Again, use the table to work out the most likely location of the transmitter. The snag here is that the transmitter might start moving after you start moving, so you have to take a series of measurements at each location. If the source appears to have started moving, stay still and use the first method.
I was recently shopping for an ADSL connection in the UK, and only one of the service agreements I looked at contained any mention of banned services or bandwidth consumption. That was for a special low-rate "no peer to peer" package, and the same company offered an unrestricted package for a higher price. It looks like broadband providers in the UK are starting to recognize that file sharing is the killer app for broadband, and many customers won't sign up for a service that restricts peer-to-peer downloading.
Of course, you can still find yourself with a high contention ratio that limits the usefulness of your connection for large downloads. I think it's only a matter of time before we see a P2P network that implements its own transport layer on top of UDP or raw sockets, avoiding TCP congestion control and allowing file sharers to squeeze more bandwidth out of their connections at their neighbours' expense.
The only problem with downloading large files in the background is that it continues to use bandwidth during the odd moments that you're using your connection for more urgent tasks, like posting to Slashdot.;-)
What I would like to see in the Linux kernel is variable priority for TCP streams, so you could leave a large download running and it would use the full bandwidth of the connection until a higher-priority stream (eg from the web browser) interrupted it. Currently all TCP streams have equal priority, so the web browser will "back off" just as much as the p2p software when the connection becomes congested.
It's not necessary to change the TCP protocol, since all the decisions take place on the local machine. Just add a system call for reducing the priority of a socket (like the nice() system call for CPU priority). As with processes, the scheduler would recognise "interactive" sockets (small, frequent reads and writes) and "batch" sockets (long lifetime, large reads and writes) and would allow interactive sockets to pre-empt batch sockets for a short time.
Well, that's the hard design work done. Now all I need is someone to perform the relatively easy task of coding the damn thing.;-)
...studies of wild chimp populations haved turned up data showing that they actually do get a significant part of their protein by eating small animals. So our predatory ancestry probably goes back at least 5 million years.
That estimate is based on the false assumption that any trait present in modern chimps was also present in the common ancestor of chimps and humans. Chimps have evolved as much in the last 5 million years as we have. They may have discovered hunting as recently as we did - it's even possible that they learned to hunt by imitating humans.
The authentication request doesn't have to take the form of a separate email - it can be an extra step in the SMTP protocol, invoked after the DATA step when a message looks suspicious. (Old mail servers would never invoke the extra step, and paranoid mail servers might invoke it for all messages.)
The idea of "hash cash" postage isn't new, but I'm glad that Microsoft is getting interested, because - like it or not - there's exactly one company that can introduce a new de facto standard for email, and that company is Microsoft. It's easy to write new protocols, but without support built into Outlook, Exchange and Hotmail, any new standard is going to have a hard time catching on. However, it should be noted that Microsoft Research does a lot of work that doesn't end up being incorporated into Microsoft products.
I think you're confusing multi-hop transfers with swarm downloading (as used in Bittorrent). In MUTE, you only connect directly to people who you know and trust in real life. All of your traffic goes through one or more of these trusted peers, and they forward it to one or more of their trusted peers and so on. If you look at netstat you'll only see the IP addresses of your friends - the actual source of the file is hidden behind a chain of other nodes, and you're hidden from the source in the same way. (Of course, the source might be one of your trusted peers, but you trust your friends not to work for the RIAA/MPAA/Burmese government, right?)
Searches are flooded (as in Gnutella) so they reach every node in the network. Responses and downloads follow the quickest path discovered by the flooding process.
Problems with this approach:
* Flooding is not scalable. Every node in the network processes every search... as the network grows, the search traffic will come to dominate the download traffic and will eventually saturate the network. So you need to limit the range of searches (as in Gnutella), and if that's not enough you can use techniques like supernodes (FastTrack) and multiple random walkers (Gia) to improve scalability.
* It's hard to join the network. You need to know someone who runs a node and trusts you enough to give you the IP address. The same situation exists with illegal drugs, and in practice they are widely available, but it might take a while for the "shady underworld of anonymous networking" to establish itself.
* Packets make multiple hops in the overlay network, which adds up to a lot of internet traffic that someone has to pay for. There's a strong incentive for people to modify their nodes not to forward packets, or only to forward packets that originate or terminate at one of their trusted peers (result: no paths more than two hops long, hopelessly fragmented network). This is similar to the free riding problem in previous filesharing networks, with the added complication that free riders are hidden from their victims. I think GNUnet has a solution to this problem (a sort of micropayment system with no currency), but as far as I know MUTE doesn't.
Slashdot Mirror
Yahoo mail allows you to disable images in HTML emails for this reason, not sure about Hotmail, Outlook, Evolution etc.
But...
the downside of using a government agency for airport security rather than private companies is that the actions of government agencies fall under the equal protection clause of the 14th amendment. Far be it from me to suggest that a person's ethnic or religious background might be a good predictor of his or her tendency to hijack planes, but should that turn out to be the case, a government agency would not be able to stop people flying on the basis of a profile that included racial or religious information.
As far as I know there is nothing to stop a private company from refusing you service on any grounds it pleases, without giving a reason. If a gas station owner says "no gas for you, please leave my property" then as far as I know that's the end of the argument. No gas for you. So a private security company might be able to use information in its "profile of a terrorist" that the TSA would not be able to use for constitutional reasons. It would not be able to arrest people on the basis of the profile, but it would be able to refuse to let them through its checkpoints. (Of course there would have to be some kind of small print in the sales agreement to the effect that "the carrier or its partners (ie the security firm) may refuse to let you on the plane without giving a reason, and you will be given a refund instead" or suchlike.)
Of course it's wrong to commit rape, along with any other violent crime. But whenever psycologists start talking, I'm skeptical. They have a theory and no hard evidence.
They may have been raped themselves, and they will probably have talked to a lot of victims of rape, so I suspect they know as much about the subject as you or I.
Actually, I'm skeptical anytime anyone starts talking about "root causes", not that I think it's inherently a bad way to think, but usually that ends up as some kind of "blame the parents of the parents of the parents of the parents of the murderer" argument.
I agree, it's sloppy thinking to confuse moral responsibility with causality. However, most penal systems aim to prevent future crimes as well as punishing past crimes, so understanding the causes of a crime is still important. If a serial rapist can be given psychological treatment that makes him less likely to commit another rape, shouldn't that be used in addition to punishment? It's not a matter of passing the buck to the rapist's parents or pedophile priests or anyone else, just a pragmatic matter of reducing the chances of future crimes.
I would guess it's mostly Sun or IBM workstations in universities and engineering companies, plus maybe a few obscure browsers/spiders that don't give enough information for Google to determine the OS.
That's as stupid as saying that it's about power, not sex. By definition, both are involved.
and is a perfectly natural activity.
That has nothing to do with whether it's right. Saying that animals do something is not a moral justification for it.
Every reason depends on a set of axioms, and how do you choose the axioms? If you arrive at them by reasoning then you must have used other, deeper axioms. Eventually you have to face the fact that rationality has its limits - it can only build on your instinctive or ingrained emotional reactions, not define them. You can build wonderful structures with logic and reason, but the foundations will always be emotional.
Personally I found this realization quite liberating - I no longer felt the need to justify my feelings rationally - but it also killed any hope I had of understanding my place in the universe on an intellectual level. So what's left? You can still hope to understand the world on an intuitive level, and to reason correctly from that intuitive foundation. Or you can kill yourself - but there's no reason to do that either. ;-)
Although I agree with your point, you referred to Europeans as "snotty" twice and "snobby" twice - I wonder if you're harbouring some stereotypes of your own? ;-)
Gosh, I can't believe I never understood the elegant simplicity of capitalism before. All those millions of people in workers' movements around the world must be idiots - thank goodness you were here to introduce them to the simple truth that money is (and should be) power.
Don't like a company? Don't work for it
For most people there's no option - if work is available they have to take it, otherwise their children go hungry and they lose their home. If you live in a country where the welfare state provides you with the option of not taking a job because you don't like the company, it's because workers in previous generations fought against capitalists like you to create that welfare state.
such is life in a dollar-vote based system
Do you honestly think it's a good idea that political power should be based on wealth? Should one person be born with millions of times more "dollar-votes" than another? Or should we do whatever we can to limit the political influence of money and ensure that each human being, regardless of his or her income, is entitled to an equal vote? One of these principles is called capitalism, and the other is called democracy. Make your choice.
True, and that's why this story is interesting - Britain believed the US was giving serious consideration to the use of force, not just planning for every contingency.
Most of the September 11 hijackers were from Saudi Arabia, but they certainly weren't funded by the Saudi government - Al Qaeda is committed to the overthrow of the Saudi royal family. If the US government avoids talking about Saudi Arabia it's not because their money supported the hijackers, but because their support for a fundamentalist dictatorship in the Middle East doesn't fit well with their supposed goal of democratizing the region.
No it doesn't. The article says that "76% of active web surfers" use non-browser applications. It doesn't say they don't use a browser. Admittedly the headline implies that, but the phrase "web surfers" indicates the opposite.
What would "75% of web connections do not use a browser" even mean? Do you think 75% of people are surfing the web with wget?
Not true. The outline of a circle is one-dimensional because you can describe any point on it with a single Cartesian coordinate.
Perhaps we'll never be completely rid of them, but with the correct application of technology we could significantly reduce their numbers. Why do you think engineers get so excited about the possibility of sending people to Mars?
It's easy, pronounce it like ";-p" but with your tongue further back in your throat.
First, the British government isn't constitutional in the same sense as the US government - there's no single document called "the British constitution". The founders of the US followed the European rationalist tradition: decide how the country should be run, write it down and embalm it for all time. (Until you change your mind - France has had five constitutions in 200 years.) In contrast, Britain's constitution follows the empirical tradition: if it ain't broke, don't fix it; when it breaks, patch it. So the British constitution is a messy tangle of legislation, common law and long-standing conventions, developed over time in a piecemeal fashion. Sort of a "release early, release often" approach to constitutional law. If the British constitution is Linux then the US constitution is Mach. (And the Magna Carta is Unix, the European Convention on Human Rights is the BSD networking stack, and the Copyright, Designs and Patents Act was written by SCO. Enough of that analogy.)
The book Systemantics, reviewed on Slashdot recently, claims that loosely-coupled systems developed in a piecemeal fashion are more stable than well-designed, tightly-coupled systems. I don't know if that's true of constitutions, but Britain has had a relatively peaceful (if slow) development from feudalism to near-democracy. Compared with almost any other country on Earth that's remarkably stable - even Belgium had a revolution.
Second, I think you're wide of the mark when you say that homage is paid to archaic traditions. British people are (in my experience) rather skeptical and cynical compared to Americans. If we tolerate archaic institutions it probably has more to do with suspicion of anyone who wants to rebuild the country in his own image (*cough*Blair*cough*) than with veneration of the past. When I visit the US I'm struck by the number of flags on display and the generally jingoistic atmosphere (and not just in the last two years). Many people seem to treat the US constitution as a sacred text, so I wonder whether there isn't more homage paid to archaic institutions in the US than in Britain (although the institutions are somewhat less archaic).
- Ernest GellnerAs a geek I'm worried by the increasing use of technological evidence in court cases. Geeks understand that technology fails, and more importantly that it can be hacked - information is not necessarily complete or correct just because it came from a computer or a camera. But most people attribute more significance to technological evidence than human evidence. Admittedly, cameras and computers don't lie of their own volition, but they can be made to lie by their operators. Evidence from a computer is only as valid as the word of the sysadmin, but is a jury more likely to believe a greasy hacker or a pristine computer printout?
Even if the information has not been tampered with, the use of high technology lends false authority to the entire chain of evidence. For example, it's well known that the odds of a false positive in a DNA comparison are billions to one. In the minds of many jurors, this will translate as "the odds that the accused is not guilty are billions to one". But in fact the odds are determined by multiplying all the probabilities in the chain of evidence - for example, what are the odds that a police officer substituted a DNA sample taken from the crime scene with one taken from the defendant? What are the odds that the defendant left DNA at the crime scene at some other time? What are the odds that the lab mixed up two samples? These are all much more likely than an error in the test itself, but because of our inability to estimate odds the strong link in the chain seems to lend strength to the weaker links, and the evidence seems unassailable.
To get back to surveillance, the best take I've heard on this issue came from a researcher in wearable computing (sorry, I've lost the link) who suggested that since we can't stop those in power from using surveillance technology, our best hope is to spread it as widely as possible (physically as well as socially). If everyone has a video camera and uses it all the time, it becomes hard to forge or misrepresent video evidence because you never know what conflicting evidence might exist. Likewise if everyone can find out where you've driven you have less to fear than if only the FBI can find out that information, because there's less chance that they'll try to misuse it.
I had no idea traffic shaping in Linux was so sophisticated. Looks like you can use setsockopt() to set the socket's priority via the TOS field. Thanks for the link!
Sure, if a long enough pause occurred between two sets of measurements you'd have to analyze them separately.
First, divide the building up into zones. Where possible, the boundaries between zones should follow "natural" boundaries that are likely to attenuate radio signals, like walls and ceilings, but if you have to cover large open spaces then you might have to draw some fairly arbitrary boundaries.
Put your receiver in the first zone and visit each of the other zones with your transmitter, recording the signal strength from each one. Then move the receiver to the second zone and repeat, and so on. Normalize the data - different hardware will produce different signals, so we're only interested in relative levels. For N zones, this gives you an NxN table of relative signal strengths between zones.
When you want to determine the location of a transmitter, first check whether the signal strength varies significantly over time. If it does, the source is probably moving. Stay in your current location (let's call it zone M) and make frequent measurements. By assuming that consecutive measurements come from the same zone or neighbouring zones (no teleporting), you can use the Mth row of your table to guess the source's location.
If the transmitter's signal strength remains constant it is probably still, so move around the building and take measurements from several zones. Again, use the table to work out the most likely location of the transmitter. The snag here is that the transmitter might start moving after you start moving, so you have to take a series of measurements at each location. If the source appears to have started moving, stay still and use the first method.
Of course, you can still find yourself with a high contention ratio that limits the usefulness of your connection for large downloads. I think it's only a matter of time before we see a P2P network that implements its own transport layer on top of UDP or raw sockets, avoiding TCP congestion control and allowing file sharers to squeeze more bandwidth out of their connections at their neighbours' expense.
What I would like to see in the Linux kernel is variable priority for TCP streams, so you could leave a large download running and it would use the full bandwidth of the connection until a higher-priority stream (eg from the web browser) interrupted it. Currently all TCP streams have equal priority, so the web browser will "back off" just as much as the p2p software when the connection becomes congested.
It's not necessary to change the TCP protocol, since all the decisions take place on the local machine. Just add a system call for reducing the priority of a socket (like the nice() system call for CPU priority). As with processes, the scheduler would recognise "interactive" sockets (small, frequent reads and writes) and "batch" sockets (long lifetime, large reads and writes) and would allow interactive sockets to pre-empt batch sockets for a short time.
Well, that's the hard design work done. Now all I need is someone to perform the relatively easy task of coding the damn thing. ;-)
That estimate is based on the false assumption that any trait present in modern chimps was also present in the common ancestor of chimps and humans. Chimps have evolved as much in the last 5 million years as we have. They may have discovered hunting as recently as we did - it's even possible that they learned to hunt by imitating humans.
Haven't used Mozilla recently, have you?
The idea of "hash cash" postage isn't new, but I'm glad that Microsoft is getting interested, because - like it or not - there's exactly one company that can introduce a new de facto standard for email, and that company is Microsoft. It's easy to write new protocols, but without support built into Outlook, Exchange and Hotmail, any new standard is going to have a hard time catching on. However, it should be noted that Microsoft Research does a lot of work that doesn't end up being incorporated into Microsoft products.
I think you're confusing multi-hop transfers with swarm downloading (as used in Bittorrent). In MUTE, you only connect directly to people who you know and trust in real life. All of your traffic goes through one or more of these trusted peers, and they forward it to one or more of their trusted peers and so on. If you look at netstat you'll only see the IP addresses of your friends - the actual source of the file is hidden behind a chain of other nodes, and you're hidden from the source in the same way. (Of course, the source might be one of your trusted peers, but you trust your friends not to work for the RIAA/MPAA/Burmese government, right?)
Searches are flooded (as in Gnutella) so they reach every node in the network. Responses and downloads follow the quickest path discovered by the flooding process.
Problems with this approach:
* Flooding is not scalable. Every node in the network processes every search... as the network grows, the search traffic will come to dominate the download traffic and will eventually saturate the network. So you need to limit the range of searches (as in Gnutella), and if that's not enough you can use techniques like supernodes (FastTrack) and multiple random walkers (Gia) to improve scalability.
* It's hard to join the network. You need to know someone who runs a node and trusts you enough to give you the IP address. The same situation exists with illegal drugs, and in practice they are widely available, but it might take a while for the "shady underworld of anonymous networking" to establish itself.
* Packets make multiple hops in the overlay network, which adds up to a lot of internet traffic that someone has to pay for. There's a strong incentive for people to modify their nodes not to forward packets, or only to forward packets that originate or terminate at one of their trusted peers (result: no paths more than two hops long, hopelessly fragmented network). This is similar to the free riding problem in previous filesharing networks, with the added complication that free riders are hidden from their victims. I think GNUnet has a solution to this problem (a sort of micropayment system with no currency), but as far as I know MUTE doesn't.